Back in November, I attended the Agile Testing and Behaviour-Driven Development Exchange (BDDX) at Skills Matter in London. Here is my very delayed write-up of the event. After a mini nightmare catching a train to Skills Matter (I watched two trains leave because they were too full and the platform was jam-packed. I’m glad I don’t have to deal with the Tube on a daily basis.), I arrived in plenty of time for an invigorating cup of coffee and an amazing carrot cake muffin. After a small spot of networking, we started promptly at 10am with a keynote from Dan North on the subject of Deliberate Discovery. This is the first talk from Dan North I have attended, and I have to say I was very impressed. Dan is an engaging and persuasive speaker, and his talk dealt effectively with the concept of second order ignorance, or what Donald Rumsfeld famously termed “unknown unknowns”. Dan’s key thesis was that even if you were to make a list of all the things that were likely to be holding your project up at any given moment, the largest, most important thing would not be on that list, because you simply don’t even know that you don’t know about it. Deliberate discovery provides some techniques for coping with the pitfalls of the usual planning process, such as positive and confirmation biases, as well as helping you realise that it’s ok to have unknown unknowns. Dan’s blog post on the subject covers much of the same material as his keynote, so between that and my notes you should have something approaching 100% coverage! </badGeekJoke>
Lunch was bracketed by a a quartet of case studies at varying levels of interest and use. The first, covering usage of SpecFlow in building a new product (a C# implementation of the Gherkin specification DSL), could feed back directly into my own work and projects, and I think some useful information was covered here. The second, from AutoTrader, was an amusing romp through adopting Agile practices in a PRINCE2/Waterfall environment, from a tester’s perspective. Much of what was covered in this talk would have been very useful to the guys at my last job. There was also a demo of a new product from ThoughtWorks called Twist, which includes some really powerful features, such as ReSharper/Eclipse-like refactoring of acceptance tests. Currently Twist only supports Java products, but .NET and Ruby support is on the roadmap. I urge you to check out this product, it looks amazingly powerful.
The two post-lunch case studies were the low point of the day for me, particularly the data warehousing talk which seemed too specific to that company’s scenario to be of general use. It was also full of data warehousing jargon like ETL, OLAP Cubs, SSIS, etc., that just went right over my head. The talk from the guys at Klarna, however, was good. They have a very interesting business model whereby they act as a middle man in eCommerce transactions: the customer pays nothing until the goods have arrived (Klarna pay the merchant in the meantime). This was borne out of a case in Sweden a while back where an online retailer was making a bunch of money selling digital cameras; one Christmas they realised they could make more money if they didn’t ship the goods, and a whole bunch of customers lost a whole bunch of money. The company was founded by a couple of business guys and a techie who built the service in Erlang in one month. In the last 5 years, their revenue has jumped nearly 14000%, the monthly income for October 2010 exceeded that for the whole of 2008, they have expanded to 380 employees, and now operate in 6 countries. That kind of rapid growth and expansion brings a bunch of problems, not least that of getting the new starters up to speed on the product as quickly as possible.
Overall, the day was very interesting. One or two of the case study talks in the middle of the day could have been a bit livelier, and might have benefitted from being broken up more by conceptual talks.